Indigenous Knowledge as a sole data source in habitat selection functions

Abstract

While Indigenous Knowledge (IK) contains a wealth of information on the behavior and habitat use of species, it is rarely included in the species-habitat models frequently used by Western species management authorities. As decisions from these authorities can limit access to species that are important culturally and for subsistence, exclusion of IK in conservation and management frameworks can negatively impact both species and Indigenous communities. In partnership with Iñupiat hunters, we developed methods to statistically characterize IK of species-habitat relationships and developed models that rely solely on IK to identify species habitat use and important areas. We provide methods for different types of IK documentation and for dynamic habitat types (e.g., ice concentration). We apply the method to ringed seals (natchiq in Iñupiaq) in Alaskan waters, a stock for which the designated critical habitat has been debated in part due to minimal inclusion of IK. Our work demonstrates that IK of species-habitat relationships, with the inclusion of dynamic habitat types, expands on existing mapping approaches and provides another method to identify species habitat use and important areas. The results of this work provide a straightforward and meaningful approach to include IK in species management, especially through comanagement processes.

Keywords: Arctic; Indigenous Knowledge; habitat modeling; ringed seals; species conservation.

Fig. 1.

Study area around Utqiaġvik, Alaska (Inset map shows Alaska, United States; the blue rectangle shows the study area; Utqiaġvik, Alaska is shown with the yellow triangle).

Fig. 2.

Methods process map of (A) the documentation of IK, (B) characterization of relationship between IK probability of species presence and habitat covariates, (C) application of beta regression to quantitative covariates, (D) mapping of IK probability of presence for review, (E) review and correction of the IK probabilities of presence, (F) using β estimates in a logistic regression equation to estimate (G) the final probability of species presence, and a return to (E) reviewing the final estimate of probability of species presence for any needed corrections. If corrections are required a return to (A) and the entire process may be needed. The size of the circles in (A) represents the proportion of hunters that indicated that species–habitat association and inform the probability values in (B).

Fig. 3.

Methods process map including the documentation of Indigenous Knowledge (IK), (A) identification of IK type, (B) characterization of relationship between IK probability of species presence and habitat covariates, (C) mapping of IK probability of species presence. The size of the circles in (A) represents the proportion of hunters that indicated that species–habitat association and inform the probability values in (B).

Fig. 4.

IK probability of ringed seal presence associated with habitat covariates and final predicted habitat selection. (A) IK probability of ringed seals presence associated with distance from shore. Yellow indicates high probability of presence and blue indicates low probability of presence. (B) IK probability of ringed seal presence associated with distance from shore with decreased probability within Admiralty Bay. Dark blue indicates high probability of presence areas, light blue indicates low probability of presence, Admiralty Bay is enclosed by the polygon shown in blue. (C) IK probability of ringed seal presence associated with ice concentration for July 15, 2015. White indicates low probability of presence and dark purple indicates high probability of presence. (D) IK probability of ringed seal presence associated with currents. Light green indicates areas of higher probability of presence, dark green indicates areas of lower probability of presence. (E) Final IK probability of ringed seal presence for July 15, 2015. White indicates areas of high probability of presence, dark blue indicates areas of low probability of presence.